Cancer1 is a leading cause of death in men and women throughout the world. In the United States alone, over 1 million new cases are diagnosed each year, and over 0.5 million deaths are reported annually (Landis et al., 1998). Historically, tumors are grouped and treated, based in part by the tissues in which they arise, e.g.—breast cancer, colon cancer, and lung cancer, and the like. Yet, within lung cancer, for example, it is well recognized that these tumors are a very heterogeneous group of neoplasms. This is also true for tumors arising in other tissues. In part, because of this heterogeneity, there are complex and inconsistent classification schemes which are used for human tumors. Previous attempts to treat cancer have been hampered by: 1) the arbitrary classification of tumors arising within given tissues, and 2) by using microscopic methods based on how these tumors look (histological classification). Although existing classifications for various tumor types have some prognostic value, almost all of the classifications fail to predict responsiveness to treatments and likelihood of cure or disease course. Improved classification schemes based on the biological constitution of these neoplasms is required to significantly alter the survival statistics of humans who have cancer. One approach to solving these problems is to locate molecules specific to tumors, preferably antigens in molecules that are markers for cancer cells. (A “marker” is defined herein as any property which can be used to distinguish cancer from normal tissues and from other disease states.) The markers' presence is then a basis for classification. Terminology used herein is as follows, “cancer is a malignant tumor, wherein a “tumor” is an abnormal mass of tissue, that need not be malignant. “Neoplasm” us a form f new growth.”
Monoclonal antibodies (MCAs) prepared by somatic cell hybridization techniques, usually in mice, are useful molecular probes for the detection and discrimination of cellular antigens, and therefore have great potential for detecting cancer associated antigens. These antibodies bind to specific antigens and the binding is detectable by well known methods. When binding occurs, the inference is made that a specific antigen is present. Those cancer associated antigens which are exposed to the cell surface or found in the cancer mass, are molecular targets for the immune systems (including host antibodies) of the host. Recent findings suggest that cancer patients who have antibodies against their tumors, do better than those who do not mount this type of immune response (Livingston, et al., 1994). Therefore, natural, induced, or administered antibodies are a promising therapeutic approach.
The humanization of non-human MCAs (the process by which non-human MCA reactive sites are shuttled into cloned human antibodies and expressed) results in reduced immunogenicity of the foreign antibodies without the loss of their specific binding in in vivo and in ex vivo applications. MCAs can be used as in vivo imaging agents diagnostic tests, and for therapy (Radosevich, et al. 1988, 1990; Rosen, et al. 1988).
Vaccine therapy is a well established approach directed at inducing an immune response without exposure to the causative agent of a disease or condition. Many vaccines are available, for example, to stimulate a response in a host to bacterial and viral agents. The use of tumor associated antigens (markers) in a vaccine could prevent primary cancer occurrence, and could also provide a means to prevent recurrence of the disease.
Gene therapy is a means by which the genetic make-up of cells is modified to express the gene of interest. There are many forms of gene therapy including: gene replacement, antisense suppression therapy, and surrogate gene expression. Discovering genes encoding cancer-associated, preferably cancer-specific antigens (markers) opens the door to genetic intervention against cancer cell proliferation. The accurate and consistent use of a cancer marker to differentiate cancerous from normal tissue, not only has diagnostic potential, but is also desirable for treatment and prognosis. Therefore, such markers have been sought.
Recent studies have shown that the enzyme encoding human aspartyl beta-hydroxylase (HAAH) is overexpressed in some human adenocarcinoma cell lines, and in primary hepatocellular cancers, therefore could be a marker. The gene said to encode HAAH has been cloned and sequenced (Gronke, et al., 1989, 1990; Wang, et al, 1991; Jia, et al., 1992, 1994; Korioth, et al., 1994; Lavaissiere, et al., 1996). However, little is known about HAAH expression in human tumors in general (Lavaissiere, et al., 1996).
The study of the HAAH enzyme grew out of the study of its bovine counterpart (Gronke, et al., 1989, 1990; Wang, et al., 1991; Jia, et al., 1992). Bovine aspartyl beta-hydroxylase is an intracellular, glycosylated protein, localized in the rough endoolasmic reticulum. The protein has been reported to have three major species of molecules; a 85 kilodalton form, and two active forms with molecular weights of 56 and 52 kilodaltons respectively (Lavaissiere, et al., 1996).
Using standard biochemical methods, bovine aspartyl beta-hydroxylase (bAAH) has been purified and characterized (Gronke, et al. 1990; Wang, et al., 1991). The activity of the enzyme has been shown to be correlated with the 52 and 56 kilodalton species which were purified. Immunologically, a related higher molecular weight form (85-90 kilodalton) was also observed. As part of the purification, bAAH is bound to Con A sepharose, which is consistent with the conclusion that the enzyme is glycosylated. (Subsequent reports on the DNA sequence show three possible glycosylation sites, with one site being very close to the known active enzyme domain.) The protein is very acidic in nature, and a detergent is not required to solubilize the active fraction. The active enzyme site is dependent from the biochemically isolated bovine protein (bAAH) on the presence of histidine at position 675 (Jia, et al., 1994).
A partial amino acid sequence was obtained for HAAH. DNA probes (a DNA probe is a molecule having a nucleotide sequence that is capable of binding to a specified nucleotide sequence under certain conditions) deduced from this amino acid sequence was used to screen a bovine cDNA library (Jia, et al., 1992). (A cDNA library contains the sections of DNA that encode for gene products, e.g. peptides as opposed to genomic DNA). Several overlapping cDNA sequences in the library contained a 764 amino acid open reading frame (ORF) sequence which will be expected to encode an 85 kilodalton protein. Also present in this ORF sequence were two other possible start codons, that is, locations at which encoding begins. The most 3′ start codon was preceded by a ribosome binding site. Translation of the clone having this sequence resulted in a protein that was about 85 kilodaltons. Antiserum was raised to the membrane fraction of human MG-63 calls and was used to immunoscreen a cDNA library made from MG-63 cells. Data on one clone was reported which could encode a 757 amino acid protein, and, by sequence analysis, was found to have strong N-terminal homology with bAAH (Korioth, et al., 1994). When this clone was used in an in vitro translation system (an artificial cocktail of normal cell cytoplasm used to convert mRNA into protein), a 56 kilodalton protein was produced. It was suggested that this was due to posttranslational cleavage.
The HAAH enzyme is responsible for the modification of specific aspartic acid residues within the epidermal growth factor-like domains of proteins. It has been hypothesized that these modified aspartic acid residues allow the epidermal growth factor-like domains to become calcium binding domains. (Gronke, et al., 1989, 1990; Wang, et al., 1991; Jia, et al., 1992, 1994; Korioth, et al., 1994; Lavaissiere, et al., 1996).
An enzyme related to HAAH, aspartyl beta-hydroxylase (AAH), was first studied because it specifically modified select aspartic acid or asparagine residues in a group of biologically important proteins including the vitamin K-dependent coagulation factors VII, IX, and X. Other proteins like C, S, and Z also have this modification (Gronke, et al, 1989, 1990; Wang, et al., 1991; Jia, et al., 1992, 1994; Korioth, et al., 1994; Lavaissiere, et al., 1996). Aspartic acid and asparagine residues have been shown to be modified by HAAH in proteins containing epidermal growth factor-like domains. The function of the beta-hydroxyaspartic and beta-hydroxyasparagine residues is unknown, however, it has been speculated that this modification is required for calcium binding in the epidermal growth factor EGF-like domains of selected proteins.
Antibodies were raised to human hepatocellular carcinoma FOCUS cells (Lavaissiere, et al., 1990). One MCA reacted with an antigen that was highly expressed in hepatocellular carcinomas (Lavaissiere, et al., 1996). Immunoscreening using this antibody and a lambda gt11 HepG2 library resulted in the isolation of a partial cDNA, which was subsequently used to isolate a larger clone.
A human adenocarcinoma cell line designated A549 was reported as having very high levels of HAAH activity (Lavaissiere, et al., 1996). A mouse monoclonal antibody designated MCA 44-3A6 (U.S. Pat. No. 4,816,402) was produced against the human adenocarcinoma cell line A549 (ATCC accession number CCL 185) (Radosevich, et al., 1985). The antibody recognized a cell surface, non-glycosylated antigenic protein having an estimated apparent molecular weight of 40 kDa).
The antigen was expressed by A549 cells, and was found to be a good adenocarcinoma marker; that is, it was frequently expressed by cancers which looked like adenocarcinomas when examined histologically (Radosevich, et al., 1990a; Lee, at al., 1985). MCA 44-3A6 is unique in that it is the first monoclonal antibody which has this binding specificity. The results from an International Workshop for Lung cancer confirmed other related published findings on MCA 44-3A6 (Stahel, 1994).
The antibody designated MCA 44-3A6 has clinical utility because it differentiates antigens associated with adenocarcinomas. The normal and fetal tissue distribution of the antigen is restricted to some glandular tissues (Radosevich, et al., 1991). Detection can occur on formalin fixed-paraffin embedded tissue (Radosevich, et al., 1985, 1988, 1990a, 1990b; Lee, et al., 1985, 1986; Piehl, et a. 1988; Combs, et al., 1988b, 1988c; Banner, et al., 1985). The antibody has a restricted binding pattern within human pulmonary tumors (Lee, et al., 1985; Banner, et al., 1985; Radosevich, et al., 1990a, 1990b).
In a study of over two hundred pulmonary cancers, MCA 44-3A6 was found to react with all of the adenocarcinomas tested, many of the large cell carcinomas, as well as with subsets of intermediate neuroendocrine small cell lung cancers, well-differentiated neuroendocrine small cell carcinomas, carcinoids, but not mesotheliomas. MCA 44-3A6 does not react with squamous cell carcinoma, bronchioloalveolar carcinoma, or small cell carcinoma (Lee, et al., 1985). MCA 44-3A6 is useful in distinguishing adenocarcinomas that are metastatic to the pleura from mesothelioma (Lee, et al., 1986). The antibody has selected reactivity among adenocarcinomas and in large cell carcinomas (Piehl, et al., 1988; Radosevich, et al., 1990b).
In a study of over 40 cases of lung cancer comparing cytological and histological findings, MCA 44-3A6 was found to be useful in cytological diagnosis and was consistent with the histological finding (Banner, et al., 1985). Histology is the study of tissues (which are made of cells). Cytology is the sturdy of cells which have been removed from the organizational context which is commonly referred to as tissue. Cells removed from tissues do not always behave the same as if they were in the tissue from which they were derived. Fortunately, the antigen detected by MCA 44-3A6 expressed in adenocarcinoma cells in tissue behaves in the same ways as adenocarcinoma cells removed from tissues. This is a very diagnostically important characteristic. Similar correlations using cytologically prepared cell blocks of pulmonary carcinomas, as well as ACs presenting in body fluids from other sites throughout the body were demonstrated (Lee, et al., 1985; Spagnolo, et al., 1991; Combs, et al:, 1988c). Also, MCA 44-3A6 binds to adenocarcinomas from sites other than lung cancer. The expression of the antigen in primary and metastatic lesions was also reported (Combs, et al., 1988a). The utility of the MCA antibody in differentiating cancer from benign lesions in human breast tissue was also noted (Duda, et al., 1991).
The cellular localization of the antigen detected by MCA 44-3A6 W2s determined. By using live cell radioimmunoassays (a radioactive antibody test directed at determining binding of the antibody to live cells), immunofluorescence, and live cell fluorescence activated cell sorter (FACS) analysis, the antigen detected by MCA 44-3A6, was shown to be on the outside surface of the cell (Radosevich, et al., 1985). Additional studies using immunogold-electron microscopy and FACS analysis have demonstrated that this antigen is non-modulated (that is not internalized by the cancer cell when bound by an antibody), is expressed on the extracellular surface of the plasma membrane, and is not cell cycle specific that is, the cell makes protein all the time it is going through the process of cell replication, and also when it is not dividing (Radosevich, et al., 1991). The antigen is not found in the serum of normal or tumor bearing patients, and is not shed into the culture media by positive cell lines (that is, cancer cells are known to bleb off portions of their cell membranes and release them into the surrounding fluid.) (Radosevich, et al., 1985). Recently 3 of 27 randomly tested adenocarcinoma patients were found to have naturally occurring antibodies to the antigen. In addition, radiolabeled MCA 44-3A6 was used to localize A549 tumors growing in nude mice. A douxorubicin immunoconjugate MCA 44-3A6 is selectively toxic in vitro (Sinkule, et al., 1991).
Determination of the nucleotide and amino acid sequences of the antigen detected by MCA 44-3A6 would enhance the usefulness of this antigen in cancer diagnosis, treatment and prevention.